ENSEMBLE FAST ELLIPSE FITTING TO OPTIMIZE THE AUTOMATIC MEASUREMENT OF FETAL HEAD CIRCUMFERENCE ON ULTRASONOGRAPHIC IMAGES

• 主要作者: Tara Avalokita, Devi
• 格式: Final Project
• 語言: Indonesia
• 在線閱讀: https://digilib.itb.ac.id/gdl/view/49320
• 機構: Institut Teknologi Bandung

Gestational age (GA) monitoring from fetal ultrasound imaging is one method to observe pre-birth risk factors and to prepare early treatment for neonatal problems. There are several parameters in the ultrasound image that can be used to estimate GA, one of which is head circumference (HC). Automatic fetal biometry measurements have been developed to reduce the variation caused by both internal and external observer factors, increase work efficiency by reducing examination time and the number of steps required in standard fetal measurements, and assisting less experienced observers. However, variations in the appearance of the fetal head on ultrasound images is making it difficult to design one general solution to measure HC automatically. This final project aims to design an algorithm to calculate HC on ultrasound images automatically and can be implemented to estimate GA. The algorithm consisted of pre-processing step to improve the selectivitiy of pixels to be processed by the ellipse-fitting algorithm, ElliFit. The algorithm is performed on 844 fetal head ultrasound images from second trimester and third trimester, with the average dice similarity coefficient of 95,69% ± 5,49%, Hausdorff distance of 20,73 pixels ± 24,18 pixels or -3,16 mm ± 4,65 mm, a difference in fetal HC of - 2,9 mm ± 11,78 mm, and an absolute difference in fetal HC of 5,86 mm ± 10,56 mm. The results demonstrated that the algorithm performed comparably to other systems published in literature.